4-acceleration

Back to a particle in general motion:

Denition 1 The 4-vector A = dV=d with components (ct; x; y; z), where
=
d=d , is called the 4-acceleration.
Recall that dt=d = . So
A = ddt ddt X :
To make this explicit, we use:
p
Lemma 1 Put (v) = 1= 1 , v =c then
2 2

d = v dv : 3

dt c dt 2

Proof: dp 1 v dv 1
dt 1 , v =c c dt (1 , v =c ) = :
2
= 2 2 2 2 3 2

2
Thus
(ct; x; y; z) = dd ( (v)(c; v))
= (v) ddt ( (c; v))
= v ddvt ( (c; v)) + (0; a)
4 2

where a = d r=dt (components of acceleration measured in the ICS).

2 2

Denition 2 The instantaneous rest frame (IRF) is the frame in which v = 0.

In the IRF v = 0 so that = 1 and V and A have components (c; 0; 0; 0) and (0; a).
We deduce:
(1) g(A; A) = ,a where a is the acceleration measured in the IRF (i.e. accelera-
2

tion felt).
(2) g(A; V ) = 0 (this can also be seen by dierentiating g(V; V ) = c .2

1
Constant Acceleration
Consider a particle moving along the x-axis with constant acceleration a measured
in its rest frame at each point.
V has components (ct_; x;_ 0; 0), and A has components (ct; x; 0; 0) where
= d=d .
We have
c = g(V; V ) = c t_ , x_ ; ,a = g(A; A) = c t , x
2 2 2 2 2 2 2 2

Dierentiate the rst equation to get c t_t = x_ x and eliminate x and x_ from 2nd:
2

c t_ t = c (t_ , 1)(c t + a )
4 2 2 2 2 2 2 2

so t = ac (t_ , 1) = which integrates to give

2 1 2

t_ = cosh(a=c) by choosing 0 for  , and

t = ac sinh(a=c) by choosing 0 for t, and
x = ca cosh(a=c) by choosing 0 for x.
2

Worldline is the hyperbola c t , x = ,c =a .

2 2 2 4 2

V
X
O
x

2
Displacement 4-vector from the origin is
X = ca (sinh(a=c); cosh(a=c); 0; 0)
2

and 4-velocity is V (note juj < c 8 )

V = c(cosh(a=c); sinh(a=c); 0; 0) :
Note g(X; X ) = ,c =a and g(X; V ) = 0. Thus for all  , an observer on the
4 2

spaceship at X ( ) reckons that the event O is simultaneous with X , and distance

c =a away. However, for a ' g (the accel. due to gravity), c =a is about a light-year.
2 2

The twin paradox revisited

Consider the earth at rest in the (x; t) system at x = c =a. At A, the space-ship is
2

at rest (relative to (x; t) coordinates) at x = c =a.

2

At B , the space-ship is at x = ca2 cosh(a=c).

Earth
t B

2
c /a , τ=0
O
A x

The time elapsed on earth is t = ac sinh(a=c) where  is the time elapsed on the
space-ship. Suppose a ' g, (acceleration due to gravity),  ' 10 years. Then
3
c=a ' 1 year and
t ' 21 e ' 11; 000 years.
10

We also have
x ' 12 e ' 11; 000 light-years ;
10

By piecing together trajectories such as these, you can travel as far as you like in as
short a time as you like (time in space-ship, distance measured on earth) if you are
allowed sucient acceleration.
For twin paradox, piece together 4 such trajectories:
t

Earth

22,000
O
light years x
44,000
years

Time lapse on space-ship is 40 years, whereas that on earth is 44,000 years.

4-force and relativistic Newton's law

Denition 3 The relativistic Newton law is
dP = F ;
d
where P is the 4-momentum of the particle and F is the 4-force.
4
This is only half the story; we need a law for what F should be. Later we will nd
an example in the 4-force that arises from the electromagnetic Lorentz force law.
We can also determine a 4-force from consideration of conservation of momentum:
Example: A photon rocket propels itself along the x-axis by ring n photons per
unit (proper) time  with frequency  in the rocket's rest frame along the negative
x-axis. Show that, if the initial mass is M the magnitude of the acceleration is
0

a = n~n~
M , c2 
0

Solution: Let the mass of the rocket be M ( ) and 4-velocity be U ( ). Conservation

of momentum over time interval  yields
0 = (MU ) + nN = UM + MU + nN
where N is the momentum of the photon. The given frequency implies g(U; N ) = ~ .
Thus, contracting the equation with U yields
c ddM = ,n~ ; so M = M , nc~ 
2
0 2

using U_ = A, g(U; A) = 0. However, squaring the conservation of momentum

equation yields
2
c M_ , M a = 0 ;
2 2 2

so, taking the positive solutions for a:

a = , cMddM = n~n~
2

M , c2  0

5
Maxwell's equations

We saw in lecture 2 that Maxwell's equations are

r
E = =" 0 (1)
r
B = 0 (2)
r ^ B , "  @@tE =  J
0 0 0 (3)
r ^ E + @B = 0 @t (4)
f = e(E + v ^ B) (5)
Here E is the electric eld, B the magnetic eld and  the charge density, and J the
current density and f is the force experienced by a particle of charge e due to elds
E and B (this serves as an experimental denition of E and B).
The rst 4 equations are relativistically invariant (and indeed motivated relativ-
ity) but the last is not, only true for jvj << c.

Charge and current density

The charge and current density are dened as:
Pe
Charge density: = Vlim ! V
=
0

P ev
Current density: = Vlim! V
0
=J
so , J are functions of (t; x; y; z) and the volumes V are small spheres centred at
r = (x; y; z). They are related by the continuity equation, which expresses the fact
that charges are neither created or destroyed.
For a xed volume V bounded by a surface S ,
R
(a) V dV = total charge in V .
R
(b) @V J
dS = total charge leaving V per unit time (negative charges entering
count as positive charges leaving).

6
where dS = ndS and n is the outward normal.
Consider N particles/unit volume each with charge e, velocity v at time t. Those
that cross dS between t and t + t occupy at time t + t a volume dS n
vt, so they
contribute
eN v
dSt
to the out ow. Sum over ev, to get total out ow through dS of J
dSt. Then
charge conservation implies
d Z Z Z  @ 
0 = dt dV + J
ds = @t + r
J dV :
V S V

This holds for any V . Hence we obtain the continuity equation, true for all physical
sources:
@ + r
J = 0 :
@t
The origin of Maxwell's equations
All but the 2nd term (the displacement current) in equation (6) were known be-
fore Maxwell, but in an integrated form. Without that, Maxwell's equations are
dierential forms of 4 key laws.
In terms of point charges, one has the (non-relativistic) laws that a particle with
velocity v, (jvj << c) feels a force f = e(E + v ^ B), (5), (denes E and B) and
generates elds
1. E = er=4" r , (Coulomb's law) and
0
3

2. B =  ev ^ r=4r (Biot and Savart's law).

0
3

3. Faraday's law of induction (described later).

7
 E B
e’

r r

e
v

 Both E and B fall o like 1=r . 2

 The electric force law is like gravity but force between like charges is repulsive.
 They are inconsistent with relativity (e.g. (2) cannot hold for lare v or r as
changes cannot propagate with 1 speed). Maxwell's extra term result in
relativistic equations.
Maxwell put these laws into dierential form; laws 1 and 2 are equivalent to
Maxwells equations without @ E=@t and @=@ B (the laws of electrostatics and mag-
netostatics):
r
E = =" 0

r
B = 0
r^B =  J 0

r^E = 0
The procedure is analogous to that for gravity as discussed in mods and so we recall
it next.

Gravity
Recall the two descriptions of gravity:
(I) Particles and action at a distance with force Gmm0=r : 2

8
 m Gmm’ m’
2
r
(II) Particles and elds g = ,r ( = gravitational potential and g = gravita-
tional acceleration) and force f = mg. The eld satises the eld equations
r
g = ,r  = ,4G ; r ^ g = 0
2

where (r) is the mass density at r.

Proposition 1 These formulations are equivalent.
Proof: I ) II:
For a single particle M , setting  = 0 at 1, we get
g = , GM r = ,r ; where  = , Gm
3
r r
Sum up over a mass distribution with density (r) to get
Z G(r0)
(r) = , dV 0
jr , r0j
V0

r r-r’

boundary
r’
dV’

9
Since r1=jr , r0j = ,r01=jr , r0j where r0 is the gradient wrt r0
 G  Z
r = , r jr , r0j dV 0
Z V  G  0

= r0 jr , r0j dV 0 ;
VZ
G r0dV 0 ;
0

= , jr , r0j
V0

by parts (i.e. moving r is equivalent to moving the body). So dierentiating under

the integral and using r jr , r0j, = 0:
2 1

Z
r  = , Gr 1
r0dV 0 ;
2
jr , r0j
Z V0

= Gr0 jr ,1 r0j
r0dV 0 ;

ZV 0
 1

= Gr0

j r , r 0 j r0
dV 0 ;
V Z 0

= "lim G r 1
dS0 ;
! 0 jr , r0j
S"

To obtain the last line we have applied the divergence outside a small sphere radius
" assuming  7! 0 at 1.

r
Sε
But (putting r = 0 for simplicity)
r0 r10 = ,r0r ; dS0 = , rr0 dS 0 ; r0 = "
0 0
3

10
Hence Z G
r  = " dS 0 ,! 4G(0)
2
2
S"

since the area of S" is 4" . Thus r  = 4G.

2 2
2
II ) I: Consider potential of a uniform sphere. Then by spherical symmetry, g =
g(r)^r. If the total mass is M , r
g = ,4G implies
Z Z Z
g
dS = r
gdV = ,4G dV = ,4GM
S V

where M is the mass enclosed by a sphere of radius r > a centre 0. Thus

Z
g
dS = 4r g(r) = ,4GM
2

so that g = ,GM=r . 2
2

Electrostatics
An identical argument, then, gives the equivalence of Coulomb's law with the dif-
ferential equations
r
E = =" ; r ^ E = 00

and the existence of a potential  satisfying

E = ,r ; r  = ,=" 2
0

where  is now the electric potential and  is now the charge density. The reversal
of sign yields a repulsive force for like charges. The potential is measured in volts
`energy/charge'.
Problems in electrostatics can be solved by mods maths methods, except that
we have new boundary conditions:
E.g. an earthed conductor has  = 0 (so that E = r = 0 tangent to conductor)
and  = 0 at 1.
Example: Consider an earthed conducting plate in the (x; y)-plane with charge
at (0; 0; a). What is the force between the charge and the plate?
Solution: We must nd  s.t.
11
 1. r  = 0 on z  0,
2

2.  = 0 at z = 0,
3.  ! 0 at 1,
4.  = " x2 y2e z,a 2 12 + where is smooth on z  0 including (0; 0; a).
4 0( + +( ) )

Place an image charge ,e at (0; 0; ,a) to get

!
e 1 1
 = 4" 1 ,
0 (x + y + (z , a) ) 2 (x + y + (z + a) ) 12
2 2 2 2 2 2

and we nd that there is an attractive force

f = 4" e(2a) :
2

2
0

12